Before this patch, the log might be corrupted by other threads logging
at the same time. For example, another RPC thread:
[httpworker.1] [default wallet] keypool reserve 1296
[httpworker.1] SelectCoins() best subset: Received a POST request for / from 127.0.0.1:53732
[httpworker.3] ThreadRPCServer method=getnetworkinfo user=__cookie__
[httpworker.1] 0.78125 0.1953125 0.02441406 0.00610351 0.00305175 0.00152587 total 1.01025417
Improve readability of code, simplify future scripted diff cleanup PRs, and be
more consistent with naming for GetBoolArg.
This will also be useful for replacing runtime settings type checking
with compile time checking.
-BEGIN VERIFY SCRIPT-
git grep -l GetArg | xargs sed -i 's/GetArg(\([^)]*\( [0-9]\+\|-1\|port\|BaseParams().RPCPort()\|Params().GetDefaultPort()\|_TIMEOUT\|Height\|_WORKQUEUE\|_THREADS\|_CONNECTIONS\|LIMIT\|SigOp\|Bytes\|_VERSION\|_AGE\|_CHECKS\|Checks() ? 1 : 0\|_BANTIME\|Cache\|BLOCKS\|LEVEL\|Weight\|Version\|BUFFER\|TARGET\|WEIGHT\|TXN\|TRANSACTIONS\|ADJUSTMENT\|i64\|Size\|nDefault\|_EXPIRY\|HEIGHT\|SIZE\|SNDHWM\|_TIME_MS\)\))/GetIntArg(\1)/g'
-END VERIFY SCRIPT-
Co-authored-by: Hennadii Stepanov <32963518+hebasto@users.noreply.github.com>
ee7891a0c4 doc: Remove outdated comments (Hennadii Stepanov)
Pull request description:
The first removed comment was introduced in #5288, the second one in #13503.
Both are outdated since #14336.
ACKs for top commit:
duncandean:
crACK ee7891a0
Tree-SHA512: a2d6071919e81c916bfc2178109bbc464417321bcc567ed0644448c5faea8e58cb08a7657afa1b6ffe1fb63e114a2a47b31c893e471839ba9d49a3986e68b2a7
faa9c19a4b doc: Add 23061 release notes (MarcoFalke)
faff17bbde Fix (inverse) meaning of -persistmempool (MarcoFalke)
Pull request description:
Passing `-persistmempool` is currently treated as `-nopersistmempool`
ACKs for top commit:
jnewbery:
reACK faa9c19a4b
hebasto:
ACK faa9c19a4b, I have reviewed the code and it looks OK, I agree it can be merged.
Tree-SHA512: f34a89a07745dabe340eb845b2a348b79c093e9056f7a21c17e1ba2e278177c9b4cf30e8095791fd645a7f90eb34850b2eee0c869b4f6ec02bf749c73b0e52ee
d96b000e94 Make GUI UTXO lock/unlock persistent (Samuel Dobson)
077154fe69 Add release note for lockunspent change (Samuel Dobson)
719ae927dc Update lockunspent tests for lock persistence (Samuel Dobson)
f13fc16295 Allow lockunspent to store the lock in the wallet DB (Samuel Dobson)
c52789365e Allow locked UTXOs to be store in the wallet database (Samuel Dobson)
Pull request description:
Addresses and closes#22368
As per that issue (and its predecessor #14907), there seems to be some interest in allowing unspent outputs to be locked persistently. This PR does so by adding a flag to lockunspent to store the change in the wallet database. Defaults to false, so there is no change in default behaviour.
Edit: GUI commit changes default behaviour. UTXOs locked/unlocked via the GUI are now persistent.
ACKs for top commit:
achow101:
ACK d96b000e94
kristapsk:
ACK d96b000e94
lsilva01:
Tested ACK d96b000e94 on Ubuntu 20.04
prayank23:
ACK d96b000e94
Tree-SHA512: 957a5bbfe7f763036796906ccb1598feb6c14c5975838be1ba24a198840bf59e83233165cb112cebae909b6b25bf27275a4d7fa425923ef6c788ff671d7a89a8
e148a52332 bench: fixed ubsan implicit conversion (Martin Ankerl)
da4e2f1da0 bench: various args improvements (Jon Atack)
d312fd94a1 bench: clean up includes (Jon Atack)
1f10f1663e bench: add usage description and documentation (Martin Ankerl)
d3c6f8bfa1 bench: introduce -min_time argument (Martin Ankerl)
9fef832932 bench: make EvictionProtection.* work with any number of iterations (Martin Ankerl)
153e6860e8 bench: change AddrManGood to AddrManAddThenGood (Martin Ankerl)
468b232f71 bench: remove unnecessary & incorrect multiplication in MuHashDiv (Martin Ankerl)
eed99cf272 bench: update nanobench from 4.3.4 to 4.3.6 (Martin Ankerl)
Pull request description:
This PR updates the nanobench with the latest release from upstream, v4.3.6. It fixes the missing performance counters.
Due to discussions on #22999 I have done some work that should make the benchmark results more reliable. It introduces a new flag `-min_time` that allows to run a benchmark for much longer then the default. When results are unreliable, choosing a large timeframe here should usually get repeatable results even when frequency scaling cannot be disabled. The default is now 10ms. For this to work I have changed the `AddrManGood` and `EvictionProtection` benchmarks so they work with any number of iterations.
Also, this adds more usage documentation to `bench_bitcoin -h` and I've cherry-picked two changes from #22999 authored by Jon Atack
ACKs for top commit:
jonatack:
re-ACK e148a52332
laanwj:
Code review ACK e148a52332
Tree-SHA512: 2da6de19a5c85ac234b190025e195c727546166dbb75e3f9267e667a73677ba1e29b7765877418a42b1407b65df901e0130763936525e6f1450f18f08837c40c
fa4db8671b test: Activate all regtest softforks at height 1, unless overridden (MarcoFalke)
faad1e5ffd Introduce -testactivationheight=name@height setting (MarcoFalke)
fadb2ef2fa test: Add extra_args argument to TestChain100Setup constructor (MarcoFalke)
faa46986aa test: Remove version argument from build_next_block in p2p_segwit test (MarcoFalke)
fa086ef539 test: Remove unused ~TestChain100Setup (MarcoFalke)
Pull request description:
All softforks that are active at the tip of mainnet, should also be active from genesis in regtest. Otherwise their rules might not be enforced in user testing, thus making their testing less useful.
To still allow tests to check pre-softfork rules, a runtime argument can change the activation height.
ACKs for top commit:
laanwj:
Code review ACK fa4db8671b
theStack:
re-ACK fa4db8671b
Tree-SHA512: 6397d46ff56ebc48c007a4cda633904d6ac085bc76b4ecf83097c546c7eec93ac0c44b88083b2611b9091c8d1fb8ee1e314065de078ef15e922c015de7ade8bf
There are two more cases where waste can be 0, when:
- (Fee - LTF) == -Change Cost
- (Fee - LTF) == -Excess
Adding these two conditions explicitly in the unit test will help
pin the behavior, also demonstrate waste calculation scenarios to new
readers.
It was [pointed out in #23030](https://github.com/bitcoin/bitcoin/pull/23030#issuecomment-922893367) that we might be able to get rid of our weak linking of [`getauxval()`](https://man7.org/linux/man-pages/man3/getauxval.3.html) (`HAVE_WEAK_GETAUXVAL`) entirely, with only Android being a potential holdout:
> I wonder if it's time to get rid of HAVE_WEAK_GETAUXVAL. I think it's confusing. Either we build against a C library that has this functionality, or not. We don't do this weak linking thing for any other symbols and recently got rid of the other glibc backwards compatibility stuff.
> Unless there is still a current platform that really needs it (Android?), I'd prefer to remove it from the build system, it has caused enough issues.
After looking at Android further, it would seem that given we are moving to using `std::filesystem`, which [requires NDK version 22 and later](https://github.com/android/ndk/wiki/Changelog-r22), and `getauxval` has been available in the since [API version 18](https://developer.android.com/ndk/guides/cpu-features#features_using_libcs_getauxval3), that shouldn't really be an issue. Support for API levels < 19 will be dropped with the NDK 24 release, and according to [one website](https://apilevels.com/), supporting API level 18+ will cover ~99% of devices. Note that in the CI we currently build with NDK version 22 and API level 28.
The other change in this PR is removing the include of headers for ARM intrinsics, from the check for strong `getauxval()` support in configure, as they shouldn't be needed. Including these headers also meant that the check would basically only succeed when building for ARM. This would be an issue if we remove weak linking, as we wouldn't detect `getauxval()` as supported on other platforms. Note that we also use `getauxval()` in our RNG when it's available.
I've checked that with these changes we detect support for strong `getauxval()` on Alpine (muslibc). On Linux, previously we'd be detecting support for weak getauxval(), now we detect strong support. Note that we already require glibc 2.17, and `getauxval()` was introduced in `2.16`.
This is an alternative / supersedes #23030.
This ensures that if we're going to add an action to open up
a transaction in a third-party link (block explorer) that it
is seperated into it's own section.
The text for an open third-party tx URL action
is improved by appending the host name with "Show in".
This makes it self-explanatory what the action will do.
673a5bd337 test: validation: add unittest for UpdateTip behavior (James O'Beirne)
2705570109 test: refactor: separate CreateBlock in TestChain100Setup (James O'Beirne)
298bf5d563 test: refactor: declare NoMalleation const auto (James O'Beirne)
071200993f move-only: unittest: add test/util/chainstate.h (James O'Beirne)
8f5710fd0a validation: fix CheckBlockIndex for multiple chainstates (James O'Beirne)
5a807736da validation: insert assumed-valid block index entries into candidates (James O'Beirne)
01a9b8fe71 validation: set BLOCK_ASSUMED_VALID during snapshot load (James O'Beirne)
42b2520db9 chain: add BLOCK_ASSUMED_VALID for use with assumeutxo (James O'Beirne)
b217020df7 validation: change UpdateTip for multiple chainstates (James O'Beirne)
665072a36d doc: add comment for g_best_block (James O'Beirne)
ac4051d891 refactor: remove unused assumeutxo methods (James O'Beirne)
9f6bb53935 validation: add chainman ref to CChainState (James O'Beirne)
Pull request description:
This is part of the [assumeutxo project](https://github.com/bitcoin/bitcoin/projects/11) (parent PR: #15606)
---
Modify UpdateTip and CheckBlockIndex for use with multiple chainstates. Includes a new unittest verifying `g_best_block` behavior (previously untested at the unit level) and various changes necessary for running and testing `ProcessNewBlock()`-like behavior on the background validation chainstate.
This changeset introduces a new block index `nStatus` flag called `BLOCK_ASSUMED_VALID`, and it is applied to block index entries that are beneath the UTXO snapshot base block upon snapshot load. Once each block is validated (during async background validation), the flag is removed. This allows us to avoid (ab)using `BLOCK_VALID_*` flags for snapshot chain block entries, and preserves the original meaning of those flags.
Note: this PR previously incorporated changes to `LoadBlockIndex()` and `RewindBlockIndex()` as noted in Russ' comments below, but once I generated the changes necessary to test the UpdateTip change, I decided to split this changes out into another PR due to the size of this one.
ACKs for top commit:
achow101:
ACK 673a5bd337
jonatack:
Code-review re-ACK 673a5bd337 reviewed diff, rebased to master/debug build/ran unit+functional tests
naumenkogs:
ACK 673a5bd337
fjahr:
Code review ACK 673a5bd337
ariard:
utACK 673a5bd3
ryanofsky:
Code review ACK 673a5bd337. Just linker fix and split commit changes mentioned https://github.com/bitcoin/bitcoin/pull/21526#issuecomment-921064563 since last review
benthecarman:
ACK 673a5bd337
Tree-SHA512: 0a6dc23d041b27ed9fd0ee1f3e5971b92fb1d2df2fc9b655d5dc48594235321ab1798d06de2ec55482ac3966a9ed56de8d56e9e29cae75bbe8690bafc2dda383
64e1ddd255 log: call LogPrint only once with time data samples (Martin Zumsande)
Pull request description:
When timedata samples are logged, `LogPrint()` is currently invoked multiple times on the same log entry.
This can lead to chaos in the log when other threads log concurrently, as in this example which motivated this PR:
```
2021-09-20T00:28:57Z -48 -26 -11 -8 -6 Addrman checks started: new 37053, tried 83, total 37136
2021-09-20T00:28:57Z -3 -1 -1 -1 -1 +0 | nTimeOffset = -3 (+0 minutes)
```
Fix this by building the log message in a string and logging it one `LogPrint()` call. I also changed the wording slightly so that it becomes understandable what is being logged, example:
```
2021-09-21T21:03:24Z time data samples: -43 -18 -12 -4 -1 -1 +0 +0 +268 | median offset = -1 (+0 minutes)
```
ACKs for top commit:
jnewbery:
Tested ACK 64e1ddd255
laanwj:
Tested ACK 64e1ddd255, new message lgtm
Tree-SHA512: ffb7a93166cc8fd6a39200b9e03a9d1e8e975b7ded822ccddd015f553258b991162a5cb867501f426d3ebcfef4f32f0e06e17b18e6b01486b967595d102f8379
ab27800799 log: Remove unnecessary timing logs for Callbacks bench (Douglas Chimento)
Pull request description:
Logging of Callbacks are no longer needed and records times that are not relevant for performance analysis.
resolves#23071
ACKs for top commit:
laanwj:
Thanks. re-ACK ab27800799
jonatack:
Code review ACK ab27800799
Tree-SHA512: be1ea780c4db9407a8799065a8824b9d3610abac72af5907809ed62d493d5a54e65735de45ec5fdd0edb85ef21ec6036105abe8ca00093942980f6f92e7fec50
0ef08f8bed add missing includes in policy/rbf (glozow)
c6abeb76fb make MAX_BIP125_RBF_SEQUENCE constexpr (glozow)
3cf46f6055 [doc] improve RBF documentation (glozow)
c78eb8651b [policy/refactor] pass in relay fee instead of using global (glozow)
Pull request description:
Followups to #22675 and documentation-only changes intended to clarify the code/logic concerning mempool Replace-by-Fee.
ACKs for top commit:
jnewbery:
utACK 0ef08f8bed
fanquake:
ACK 0ef08f8bed
Tree-SHA512: 6797ae758beca0c9673cb00ce85da48e9a4ac5cb5100074ca93e004cdb31d24d91a1a7721b57fc2f619addfeb4950d8caf45fee0f5b7528defbbd121eb4d271f
fa08d4cfb1 Use C++11 member initializer in CTxMemPoolEntry (MarcoFalke)
Pull request description:
This removes a bunch of boilerplate, makes the code easier to read.
Also, C++11 member initialization avoids accidental uninitialized
members.
Can be reviewed with the git option "--word-diff-regex=." or with "git
difftool --tool=meld".
ACKs for top commit:
jnewbery:
Code review ACK fa08d4cfb1
shaavan:
Code Review ACK fa08d4cfb1
Tree-SHA512: 2424861002fbcef2a3f01845662c115b973a7a5103f359305b5d9237c055eb003aa7646fc1cb30e6eaf90810d662f94cedc6f90795e30b56680f9c81f631d64b
fa45a1338a refactor: Remove unused validation includes (MarcoFalke)
Pull request description:
Unused includes will cause needless recompilation when headers are changed. Also, they pretend there are dependencies that don't exist.
Fix both by removing them.
ACKs for top commit:
laanwj:
Code review ACK fa45a1338a
theStack:
ACK fa45a1338a♻️
Tree-SHA512: 69190fd09184b75bce34ce3f315a1817e09ea32779f9ddc2d4790c89b0887b6cebd88aba66fa054c43c9183fc66202a556d982dd7034fc389a75802d8aaac83a
Can be reviewed with -W --ignore-all-space
Fixes:
* Calling ConsumeRandomLengthByteVector 4 times, when 2 is enough.
* Slow execution speed: Finalize is expensive because it invokes
division. Speed up the target by calling Finalize() at most twice per
fuzz input.
This removes a bunch of boilerplate, makes the code easier to read.
Also, C++11 member initialization avoids accidental uninitialized
members.
Can be reviewed with the git option "--word-diff-regex=." or with "git
difftool --tool=meld".
The benchmarks can now run much longer due to the minimum of 10ms or
directly with -min_time. With -min_time=20000 I could trigger two ubsan
errors in the benchmarks, which are fixed in this commit by using
unsigned type and adding "& 0xFF".
- use ALLOW_BOOL for -list arg instead of ALLOW_ANY
- touch up `-asymptote=<n1,n2,n3...>` help
- pack Args struct a bit more efficiently
- handle args in alphabetical order
When it is not easily possible to stabilize benchmark machine and code
the argument -min_time can be used to specify a minimum duration
that a benchmark should take. E.g. choose -min_time=1000 if you
are willing to wait about 1 second for each benchmark result.
The default is now set to 10ms instead of 0, which should make runs on
fast machines more stable with negligible slowdown.
Moves copying of the setup into the benchmark loop so it is possible
to run the loop for an arbitrary number of times.
The overhead due to copying the candidates inside the loop is about 3%.
Moves some of the setup into the benchmark loop so it is possible to run
the loop for an arbitrary number of times. Due to recent optimizations
in #22974 the benchmark now runs much faster, so the inner loop now calls
Good() 32 times as often to get better numbers.
Renamed the benchmark to AddrManAddThenGood because that's now what is
actually tested. To get the the number of just Good(), one needs to
subtract the benchmark result of AddrManAdd.
Introduced in #19055, MuHashDiv benchmark used to multiply with a loop
based on epochIterations. That does not do what it is supposed to do,
because epochIterations() is determined automatically from nanobench.
Also, multiplication is not needed for the algorithm (as pointed out by
a comment in #19055), so it's better to remove this loop.
